Fasteners / Joint Design

1
Fasteners / Joint Design
NSTX TF FLAG JOINT REVIEW 4/10/03
Michael Kalish
2
Stud Preload

• Maintaining the preload on the stud is critical
  for maintaining contact pressure and contact
  resistance
• Using a long narrow bolt results in a much higher
  bolt elasticity than that of the Flag (10X).
• Applied cyclical loading adds relatively small
  additional loading to the stud.
• With higher elasticity, loss in preload due to
  deflection is minimized.

3
Preload Continued

• Belleville washers are used to account for any
  unexpected yielding of bolt or copper
• While the bolt length provides adequate
  elasticity to accommodate design load scenarios
  the addition of Belleville washers prevents
  relief of the preload in the event of
  unanticipated strain
• Washer has ½ the stiffness of the bolt, for every
  .001 inch strain only 125 lbf preload is lost
  (total washer deflection .032)
• With a strain as high as .008 washers will
  prevent preload from dropping below 3,560 lbf.
  (the first .0045 drops the bolt pre-load to
  4,000 lbf then the washer takes over)
• Testing of prototype will verify that preload is
  maintained.
• A washer plate is added to spread out the
  compressive forces under the nut and minimize
  local yielding of the copper.
• Bolts to be pre-tensioned to eliminate stored
  torque

4
Bolt Characteristics

• Bolt is a a 3/8-16 stud threaded at both ends
• To increase elasticity the bolt shank diameter
  will be just slightly larger than the root
  diameter of the threads resulting in a relatively
  elastic bolt (a creep of .001 results in a loss
  of 225 lbf of preload)
• Loading
• A preload of 5,000 lbf is applied with an
  equivalent root diameter stress of 64,700 psi
• Thermal loading after ratcheting applies an
  enforced deflection of .0043 inches corresponding
  to a stress adder of 12,700 psi
• As a result of stiffening the hub structure
  additional mechanical loading is minimal so that
  almost all fatigue loading is the result of
  thermal stress
• With the 5,000 lbf preload and the thermal
  stresses applied the bolt sees a mean tensile
  stress of 71.2 ksi and a mean amplitude of 6.5
  ksi
• The ultimate tensile strength for the A286 is 145
  ksi and the yield is 100 ksi

5
Modified Goodman Diagram For A286 Stud
6
Threaded Insert

• A TapLok 3/8-16 Medium Length insert is used
  (OD into copper is .50)
• Loading
• The bolt preload of 5,000 lbf results in 11,800
  psi in shear at the outer threads of the insert
  into the copper.
• Thermal loading adds a cyclical load of 2,300 psi
• Additional mechanical loading is minimal so that
  almost all fatigue loading is the result of
  thermal stress
• Per the inspection certification the Tensile
  strength 38 kpsi equivalent to a Shear Strength
  22 kpsi
• With the 5,000 lbf preload and the thermal
  stresses applied the bolt sees a mean shear
  stress of 12.9 kpsi and a mean cyclical amplitude
  of 1.2 ksi

7
Modified Goodman Diagram for Insert in Copper
Conductor
8
Threaded Insert (cont.)

• Testing shows margins may be greater then the
  numbers indicate. The lowest pull out force
  measured 11,500 lbf equivalent to 27 kpsi shear
  strength in the copper (as compared to 22 kpsi)
• The insert was tested for both pull out strength
  and pull out strength after cycling
• Cycling (one sample) did not indicate significant
  degradation to pull out strength (fatigue sample
  pulled at 12,380 lbf)
• Further testing is planned. A mechanical
  prototype will test for maintenance of preload
  after application of a cyclical load.

9
Testing
NSTX TF FLAG JOINT REVIEW 4/10/03

• Test Setup
• Pull Tests
• Cyclical Pull Testing
• Friction Tests
• Collar Shear Tests
• E-Beam Weld Tests
• Mechanical Prototype

Michael Kalish
10
Test Setup

• MTS Hydraulic Test Stand
• Plots Load vs Deflection up to 100,000 lbf
• Provides Cyclical Testing Capabilities

11
Pull Test Setup

• Designed to test pull out strength of inserts
• Inserts installed in spare lengths of conductor
• Keenserts and two lengths of Taplok inserts tested

12
Pull Test Setup (cont.)

• Specimen in Test Fixture

13
Pull Test Results KeenSerts

• KEEN SERT 1     11,120lbs
• KEEN SERT 2     12,000lbs
• KEEN SERT 3     11,880lbs
• KEEN SERT 4     11,620lbs
• KEEN SERT 5     11,500lbs
• KEEN SERT 6     11,260lbs
• KEEN SERT 7     11,500lbs
• KEEN SERT 8     11,380lbs

14
Pull Test Results Tap-Lok

• 3/8-16 H Series Tap-Lok Inserts
• 1st Sample SUMMARY (peak force) Regular Length
  0.687, Tap Drill
• TAP LOK 1 15,260lbs
• TAP LOK 2 15,500lbs (bolt broke)
• TAP LOK 3 15,260lbs (bolt broke)
• 2nd Sample SUMMARY (peak force) Regular Length
  0.687Tap Drill Plug Tapped Only
• TAP LOK 1 16,000lbs
• TAP LOK 2 15,380lbs
• TAP LOK 3 15,760lbs
• TAP LOK 4 15,500lbs
• TAP LOK 5 15,620lbs
• 3rd Sample SUMMARY (peak force) Medium Length
  0.562, Insert Tool Only
• TAP LOK 1 12,500lbs
• TAP LOK 2 12,500lbs
• TAP LOK 3 11,500lbs
• TAP LOK 4 12,500lbs
• TAP LOK 5 11,760lbs

Displacement vs Force Curve for Tap Lok 4, 3rd
Sample Peak Force 12,500 lb
15
Cyclical Testing, Pull Out

• Using same test setup medium length Tap Lok
  insert was cycled then pulled
• 5000 cycles at 4600 lbf to 6000 lbf
• 45000 cycles at 5400 lbf to 6000 lbf
• Cycled at 1 Hz Sine Wave
• Ranges represent thermal operational loads for
  25 degC and 5 degC cyclical thermal loading
• Sample pulled after subjected to Fatigue along
  with insert installed in same conductor piece.

16
Cyclical Testing Results

• Shaded area may indicate drift or creep (first
  run exhibited a negative creep drift)
• Worst case creep of .0025(if real)
• Top Plot5,000 cycles at 4,600lbf to 6,000lbf
• Bottom Plot45,195 cycles at 5,400lbf to 6,000lbf

17
Cyclical Testing Pull Out Result

• Fatigued Insert Pullout (bottom trace) vs
  Unfatigued Insert Pullout(top trace)
• Fatigued 12,380 lbf
• Unfatigued 13,380 lbf
• Previous samples showed scatter of 11,500 lbf
  to 12,500 lbf
• Pull out strength relatively unchanged
• More testing to follow

18
Friction Test Setup

• Two horrizontal load cells measure compressive
  force provided by eight 3/8th inch bolts
• Specimens are machined and plated after each run
• Vertical load is applied to offset middle block

19
Friction Test Calibrations

• The two compression loads cells were checked
  against the MTS vertical load cell and found to
  be to within .30

20
Friction Tests Results

• Force is recorded at point where inelastic
  behavior begins
• Initial testing of unplated copper resulted in
  COF .12
• Testing of plated copper increased COF values to
  .41

21
Friction Test Results (cont.)

• Mean COF .41
• Min. Value .39
• Results consistent for varying compressive loads
• Testing to continue
• Will explore improvement of COF by changing
  surface conditions

22
Collar Shear Test

• Representative sample of collar arrangement
  fabricated to characterize elasticity and shear
  strength

23
Collar Shear Test Results

• No compressive load was applied
• Shear area approx. 11.8 sq in.
• For lower values split occurred on one side. For
  higher values on both sides simultaneously
• Separation occurred at epoxy to SS interface.
• Further testing planned for improved samples
• SUMMARY
• 1 60,000lbs, 2,540 psi
• 2 58,590lbs, 3,000 psi
• 3 43,100lbs, 1,820 psi
• 4 40,000lbs, 2,040 psi
• 5 61,900lbs, 2,610 psi
• 6 65,000lbs, 2,720 psi

24
E-Beam Weld Test Specimens

• Tensile Testing performed on 1 x .50 EBeam
  welded copper bars

25
E-Beam Weld Test Specimen Results

• SUMMARY
• 3L Peak Force 13,760lbs, 27,520 psi
• 3R Peak Force 13,260lbs
• 26,520 psi
• 4L Peak Force 13,120lbs
• 26,420 psi
• 4R Peak Force 14,380lbs
• 28,760 psi

Copper Bar Weld 4R Peak Force 14,380lbs, Peak
Displacement .199
26
E-Beam Weld Test Specimen Results

• Bars separated at weld although weld area looked
  entirely homogeneous

27
Mechanical Prototype Testing

• A prototype of the flag bolted to a section of
  conductor is in fabrication.
• Vertical loading will be applied directly to the
  flag
• This mockup will test the bolted joint for
  maintenance of preload after cycling.
• Contact resistance will be monitored in real time.